Automatic End Point Detection of Plasma Etching Process Using the Multi-Way PCA of the Whole Optical Emission Spectrum

Automatic detection algorithm is needed for the real batch process and multi-way principal component analysis is developed to analyze the OES data and extract key component that capture the endpoint signal. The traditional endpoint detection technique uses a few manually selected wavelengths in the plasma etching process, which are adequate for large open area. As the integrated circuit devices continue to shrink in geometry and increase in device density, detecting the endpoint for small open area or multi-layer presents a serious challenge to process engineers. In this paper, a high-resolution optical emission spectroscopy system is used to provide the necessary sensitivity for detecting subtle endpoint signals. In the case study, we applied this algorithm to the open data sources and real etch process, which showed more reasonable features. This end point features can be used for the improved process monitoring afterwards